The LSAC site has a complete list of all the multipliers used by each school. I used it a couple months ago to see which schools valued GPAs more than LSAT scores (at least relatively) and vice versa (fyi, berkeley valued GPA the second highest out of all the 160-something schools on the LSAC multiplier list, trailing only U of Puerto Rico, both of which value GPA almost three times more than the average school, which treats 1 point on the LSAT as equal to 0.08 GPA points).
Does anyone have any insight on why schools bother with adding a constant to the estimated AI? The only reason I can tell is to modify the perception of differences between AI. For instance, the difference between an AI of 24 as compared to 26 may seem a lot bigger than that between 94 and 96 (constant of 70). Otheriwse, the constant is worthless since the AIs can't be compared with the AIs of other schools.
A caveat for all of you who are using the spreadsheets. They assume:
1) that the median LSAT is dead center between the 75th and 25th percentile values when top schools seem to have median scores closer to the 75th percentile. Same goes for GPAs.
2) that each X-th percentile admissions index for each school is equivalent to a student with a X-th percentile LSAT and X-th percentile GPA (i.e. 50th percentile admittee = 50th percentile LSAT and 50th percentile GPA). This can differ enough to bump a prediction from one range to another depending on how the actual scores are paired up.
3) that applicant pool inflation is equal to the percentage change in scores, NOT percentage change in # of applicants (this would be ridiculous as a 6% increase in applicants -- a stat from a previous year -- would lead to a school going from a LSAT score of 160 to 169.6. Thus the best number to use here is probably 1.01, as scores go up for each school about 1 point every year. Also, you may want to change the formula used to apply the inflation multiplier as it is currently applied to not only the LSAT and GPA scores but the constant added as well. The constant should remain unaffected unless the school actually modifies the multipliers used.
4) that if you score the same as the X-th percentile index score, you have a X percent chance of getting into the school. You can still use the prediction ranges relative to each other but I don't see why 50th percentile index means 50% acceptance probability. For one thing, the percentages are at the very least skewed in favor of those considered as strongly qualified at a particular school. The reason for this is simple. Students tend to attend the higher ranking schools to which they are accepted, so you dont need to accept nearly as many 25th percentile students as you would 75th percentile students (who you would have increased competition for).
Anyone else have any thoughts?